Notes RBG Edinb. 46(1): 59-79 (1989) A classification of Abies Miller (Pinaceae) A. Farjon* & K.D. Rushforth** Previous classifications of the Abies reviewed and evaluated. A number of Abstract genus are subgeneric names are lectotypified and a list of the validly named subgeneric taxa is given. A new classification scheme is proposed in which Abies is divided into ten sections, four ofwhich are further divided into subsections. Three new subsectional names are proposed. A key to the sections and subsections is given and their morphological characteristics and biogeography discussed. Introduction Since the established genus was clearly as a separate entity within the various have been made. Pinaceae, attempts at a classification In the main these been limited restricted have either to geographical areas (e.g. Engelmann, and have used limited 1878; Mayr, 1890; Patschke, 1913) or they only a range of characters as the primary basis for the scheme (e.g. Hickel, 1906 08; Patschke, 1913; Landry, 1984). A result of this is that the classifications proposed are eitherrestricted in their application or produce inconsistent and apparently artificial alliances. The classification we offer here is based on coning and vegetative characters and referenced back to the ecology and geography of the species and is thus, in our opinion, a more natural and complete system than any proposed to date. Historical Review The genus Abies was established by Millerin 1754with A. alba 1 as the type species. In addition to A. alba, Miller used the genus for A. balsamea and for nine other species which are now uniformly treated under either Picea or Other in the nineteenth Tsuga. workers, particularly century, also included species now treated as belonging to Pseudotsuga and Keteleeria. The first classification attempt at a generic was by Spach (1842) who divided the genus Abies into five sections: Section Picea (Link) Spach included only species of Picea Link. Section Piceaster Spach contained only A. pinsapo. Section Peuce included A. alba, A. A. A. Spach nordmanniana, procera (as 2 nobilis■) A. religiosa, A. sibirica, A. balsamea, A. fraseri, A. grandis, A. , spectabilis and A. pindrow. Section Peucoides Spach contained Pseudotsuga menziesii(Mirbel) Franco. Section Micropeuce Spach contained Tsuga canadensis (L.) Carriere and T. dumosa (D. Don) Endl. *State University of Utrecht. Institute of Systematic Botany, PO Box 80.102, 3508 TC Utrecht, The Netherlands. **32 Park Lane, Fareham, Hants, POI6 7 JX, England. 'The authorities for species, subspecies and varieties are given in the proposed classification, not where the name is first mentioned in the text. Throughout, where the original author has used a name which is invalid or at a different level than here the with accepted we cite taxon under the name accepted here, the originalname in parenthesis. 59 60 NOTES RBG EDINB. 46(1) Of these sections only Piceaster and Peuce contained species of Abies as presently circumscribed. The name of section Peuce Spach is illegitimate under Article 22.1 of 1.C.8.N., since it contains the type species of the genus (A. alba). The correct name for this section is Abies sect. Abies. Carriere(1855) followed Spach in his two sections oftrue Abies, i.e. Piceaster and Peuce. Gordon in 'The used the for (1858) Pinetum’ Picea D. Don as generic name Abies. divided the into sections: He genus two Section Bracteata Gordon was circumscribed by ‘the bracteas on the cones not hidden by the scales and eitherprojecting or reflexed’. It included A. balsamea, A. bracteata, A. cephalonica, A. firma, A. fraseri, A. procera (as A. nobilis), A. nordmanniana, A. alba (as A. pectinata) and A. religiosa (as A. glaucescens). Section BrevibracteataGordon comprised ‘thosekinds with the bracteas on the cones shorter than the scales, and enclosed’. It included A. amabilis, A. concolor, A. grandis, A. sibirica (as A. pichta), A. pindrow, A. pinsapo, and A. spectabilis (as A. webbiana). The names of both of these sections are illegitimate, section Bracteata Gordon because it includes the type species of the genus(A. alba), and section because it includes which is the BrevibracteataGordon A. pinsapo type species of the earlier described section Piceaster Spach. The 1875 edition of 'The ’ Pinetum lists a few additional species but in the same two groupings. Engelmann (1878) published an arrangement for the nine species which he them into four sections: recognized as occurring in North America, dividing Section Balsamea Engelm. contained A. balsamea, A. lasiocarpa and A. fraseri (type A. balsamea). Section Grandis Engelm. consisted of A. grandis, A. concolorand A. amabilis (as A. grandis var. densiflora) (type A. grandis). Section Bracteata Engelm. contained A. bracteataand A. religiosa (type A. bracteata). A. Section NobilisEngelm. contained procera (as A. nobilis) and A. magnifica (type A. procera). for ofhis sectionsbecause Engelmann was free to use the name Bracteata one Gordon’s earlier of the use name (in Picea D. Don) was illegitimate. who made three sections: The next proposal was by Mayr (1890) consisted of umbellata Section MomiM ayr A. firma, A. homolepis var. (as A. umbellata), A. alba (as A. pectinata), A. nordmanniana,A. bracteata, A. grandis and A. magnifica. Section Pindrau Mayr contained A. homolepis, A. veitchii, A. mariesii, A. spectabilis (as A. webbiana), A. pindrow (as A. pindrau), A. amabilis, A. A. and A. procera (as nobilis), A. fraseri religiosa. Section Pichta Mayr contained A. sachalinensis, A. sibirica (as A. pichta), A. balsameaand A. lasiocarpa (as A. subalpina) All three names ofthese sections, however, are illegitimate under the current rules of1.C8N.; section MomiMayrbecause it contains the type species of the genus;section PindrauMayr because it includes the type (A. procera) ofsection inclusion of the NobilisEngelm.; and section Pichta for the A. balsamea, type species of section Balsamea Engelm. CLASSIFICATION OF ABIES 61 Sargent (1898) amended Engelmann’s section Bracteata, restricting it to A. bracteata. Kent (1900) divided the genus into three sections: Section Abies (as Euabies) with A. amabilis, A. balsamea, A. cephalonica, A. cilicica, A. concolor, A. firma, A. fraseri, A. grandis, A. homolepis, A. lasiocarpa, A. mariesii, A. nordmanniana, A. numidica, A. alba (as A. pectinata), A. pindrow, A. pinsapo, A. religiosa, A. sachalinensis, A. sibirica, A. veitchii and A. spectabilis (as A. webbiana). Section Bracteata Engelm. emend. Sarg. containing only A. bracteata. Section NobilisEngelm. with A. procera (as A. nobilis) and A. magnifica. Hickel in a series of papers (1906 08) proposed three sections: Section Pseudopicea Hickel containedA. homolepis and A. speclabilis (as A. webbianä). Section Pseudotorreya Hickel consisted only of A. bracteata. Section Elate Hickel contained A. cephalonica, A. numidica, A. cilicica, A. nordmanniana, A. grandis, A. balsamea, A. religiosa, A. alba (as A. pectinata), A. amabilis, A. mariesii, A. veitchii, A.fraseri, A. sibirica, A. sachalinensis, A. firma, A.pindrow, A. lasiocarpa (including var. arizonica A. arizonica and with A. A. as subalpina as a separate species), A. pinsapo, procera (as A. nobilis). A. magnifica and A. concolor. Hickel did not designate a type species for his section Pseudopicea, and we propose that the lectotype for this section should be A. spectabilis. The name Pseudotorreya for the next section is superfluous as it repeats section Bracteata Engelm. The name ofsection Elate is illegitimate as it contains the type species of the genus. with Patschke (1913) proposed a scheme dealing the eastern Asiatic species and sections: gave two Section Marginalis Patschke contained A. spectabilis (as A. webbiana), A. pindrow, A. delavayi, and A. recurvata. Section Centralis Patschke included two subsections: Subsection Laterales Patschke consisted of A. firma, A. fargesii, A. squamata, A. veitchii, A. mariesii, A. kawakamii and A. homolepis. SubsectionMedianaePatschke comprised A.sachalinensis, A. holophylla, A. sibirica and A. nephrolepis. Section Marginalis has by our lectotypification of section Pseudopicea Hickel with A. spectabilis become a taxonomic synonym. A. kawakamii is here proposed as the lectotype for both section Centralisand subsection Laterales. A. sachalinensis is proposed as the lectotype for subsection Medianae. Franco divided the into two (1950) genus subgenera: Subgenus Pseudotorreya Franco consisted of only A. bracteata. Subgenus Sapinus (Endl.) Franco was divided into seven sections: Section Nobilis Engelm. with A. procera and A. magnifica. Section with Oiamel Franco A. religiosa, A. hickelii, A. hickelii var. oaxacana (as A. oaxacana) and A. vejarii (lectotype A. religiosa, Liu, 1971). 62 NOTES RBG EDINB. 46(1) Section Balsamea Engelm. emend. Franco with two series: Series Grandes (Engelm.) Franco with ('?) A. mexicana(A. vejarii subsp. mexicana in Farjon, in press), A. guatemalensis, A. durangensis, A. concolor and A. grandis. (type A. grandis). Series Lasiocarpae Franco with A. lasiocarpa, A. balsamea and A. fraseri (type A. lasiocarpa). Section Pichta Franco with A. sibirica (type), A. sachalinensis, A. veitchii, A. nephrolepis. A. koreana and A. kawakamii. Section Momi Franco with three series: Series Homolepides Franco with A. mariesii, A. homolepis and A. holophylla (type A. homolepis). Series Firmae Franco with A. firma. Series Sinenses Franco with A. chensiensis, A. recurvata, A. squamata, A. forrestii var. georgei (as A. georgei). A. fargesii, A. fargesii var. sutchuenensis (as